[volunteers] brie

[Daniel Gimpelevich]

Rick Moen <rick <at> linuxmafia.com> writes:

> 
> Quoting Lisa Corsetti (svlug <at> flygirl.com):
> 
> > We have THREE machines?  Ah... I think I understand now, but
> > let's see if I am right:
> > 
> > (1) The current webserver (I am unaware of its name)
> 
> svlug.svlug.org (aka, for the moment:  "www", "svlug.org", "lists", and
> some others)

I'm sure I remember it to also be "gruyere" locally, but that seems to have 
disappeared from /etc/hosts. And is it wise for the DNS contact to be 
root at svlug.org?

> > (2) Brie
> 
> AKA, for the moment:  newwww.svlug.org

Good to know.

> > (3) The virtual machine at Linnode
> 
> "gruyere" (AKA ns1.svlug.org)

News to me.

> > As for the skinning of mailman, I can look into that once the
> > machine is set up.
> 
> After which, things like parting the Red Sea will be easy.  ;-> 
> (Legions of Mailman users will be grateful for anyone who can 
> improve its aesthetics and integration).

BTW, the listinfo page for the main list still has a .shtml link.

> > As for name-based virtual WEB servers, assuming we can set up
> > the DNS properly (where *is* the master DNS server?  is it on
> > svlug.org?)
> 
> On grueyere.

This may be in addition to the DNS daemon brie might run.

> DNS daemon package "NSD" is used, employes BIND-format zonefiles,
> hashed to improve performance and save RAM.  It's a little different,
> but not very.  The main thing is re-running the zonefile compiler
> ("zonec") after changing a zonefile, and then "nsdc restart" (IIRC) to
> HUP the daemon.
> 
> I need to write a short recipe for that, and put it somewhere central,
> probably /usr/local/ .  Which is where other sysadmin stuff on gruyere
> has been accumulating, such as my recipe for compiling an Ubuntu .deb of
> Lighthttpd, and my working files for same.
> 
> The guts of NSD are in, of course, /etc/nsd/ .
> 
> > I'm just waiting for things to be up and running and then
> > there's a few things I can do....  I sure would like, for
> > example, to have svn running rather than RCS.
> 
> Perhaps you would consider "git"?  svn is of course serviceable (and
> easy for cvs users), but "git" keeps getting more and more and more
> people backing it as a superior solution.  The latest it Ted T'so:
> http://tytso.livejournal.com/29467.html

Like CVS, but unlike RCS, svn is primarily geared toward a code release cycle. 
Like RCS but more so, git is intended to be used on any ol' data. However, git 
is also ridiculous bloatware in the sense that everyone doing anything with the 
data must have a copy of the entire repository, along with all revisions ever 
made, or at least I have never found any other way.